Conventional multi-unmanned aerial vehicle (UAV) assisted non-orthogonal multiple access (NOMA) wireless caching networks (WCNs) usually operate in a distributed and non-collaborative manner, where each UAV serves users independently without coordination or relay support. When UAVs move beyond the communication range of ground base stations (BSs), backhaul disruption occurs, leading to high user latency and limited system scalability. To address these issues, we propose a multi-layer UAV-assisted NOMA WCN architecture, where a primary UAV (PUAV) communicates with the BS and cooperates with multiple secondary UAVs (SUAVs). The PUAV not only acts as a control and coordination hub but also serves as a relay for content transmission to SUAVs when necessary. To minimize user transmission latency, we propose a joint iterative algorithm that integrates user clustering, user pairing, power allocation, and UAV deployment. First, we develop an Advanced Balanced K-Means++ (ABKM) algorithm to ensure that each cluster contains a balanced number of users and to reduce the distance between users and their serving SUAVs. Next, we derive the NOMA power allocation factor that minimizes user transmission latency, ensuring efficient resource distribution among all paired users. Furthermore, we analyze the impact of PUAV and SUAV placement on user latency and propose a two-stage particle swarm optimization (PSO)-based algorithm to iteratively optimize the deployment of all UAVs. Finally, the user pairs and power allocation are jointly optimized based on the updated deployment of all UAVs to further reduce user latency. Simulation results show that, compared with a single-layer UAV architecture and benchmark schemes, the proposed multi-layer design with joint optimization achieves lower user latency. Additionally, comparisons with the optimal power allocation search method confirm the validity of the derived NOMA power allocation factor.

The integration of large language models (LLMs) with mobile edge computing (MEC) systems presents a novel approach to enhancing vehicle-to-everything (V2X) connected autonomous driving. This study aims to address the prevalent challenges in multi-model sensor data fusion, such as latency, privacy preservation, and the need for dynamic adaptation to evolving environmental conditions, by leveraging real-time data from LiDAR sensors. We propose an LLM-based framework to improve V2X driving assistance systems’ operational efficiency, safety, and reliability, where pictures and image recognition work as integrated data from multiple sensors to train various vehicle and lane detection models. Based on the benefits of federated learning, i.e., distributed at each MEC server and optimising models accordingly, these training models can avoid the data privacy issue in V2X driving assistance implementation. The application of generated test data significantly improves the success rate of the lane detection feature and pedestrian detection, by 95% and 85%, respectively. The experiment results demonstrate that our proposed framework is effective and feasible.

To ensure the safety of both aircraft and operators in low-altitude economy (LAE) activities, precise environmental perception capabilities are essential for effective collision prevention. However, achieving accurate perception remains challenging, particularly when obstacles are camouflaged and visually blended into their surroundings, making detection difficult, even with the robust foundation model, the Segment Anything Model (SAM). Although SAM's prompt-based strategy improves its performance in the Camouflaged Object Detection (COD) task, its reliance on limited prompts introduces new challenges. Instead of manually annotating prompts, our work introduces a multimodal learning approach that utilizes a Vision-Language Model (VLM) to automatically generate mask prompts. By integrating visual and textual information, this work generates high-quality prompts that significantly enhance the performance of SAM in identifying camouflaged objects. Experimental results demonstrate that the proposed method achieves an average improvement of 13% over the baseline SAM across three COD benchmark datasets.

With the concept of low-altitude economy (LAE) being released recently, the research on ultra-low latency communication and rich computation capacity technologies to support LAE-enabled internet of things (IoT) networks has attracted interest from industry and academia. One of the key challenges is to reduce the latency of the IoT networks while guaranteeing the quality of service among all user devices (UDs). In this paper, we propose an LAE-enabled IoT network, where a UAV-carried mobile edge computing (MEC) server offers extra computation capacity to all UDs to process their computational tasks remotely. To minimise the total service delay of all UDs, which consists of the transmission delay, processing delay, queueing delay, and UAV mobility delay, we propose a deep-Q-leaning (DQN)-based optimisation algorithm by jointly optimising the task offloading decisions and communication and computation resource allocation for all the UDs in the LAE-enabled IoT network. Simulation results illustrate that our proposed algorithm achieves a much lower total service delay than the benchmarks.

In vehicular communication and mobile edge computing (MEC) networks, limited storage resources and challenges related to vehicles data security pose significant concerns. To address these issues while reducing communication latency and enhancing network security, blockchain technology is introduced. Additionally, deep reinforcement learning (DRL) is leveraged to optimize content caching strategies. By formulating a Markov decision process (MDP) model and applying the proximal policy optimization (PPO) algorithm, efficient cache management and optimal resource allocation are achieved. Simulation results demonstrate that the proposed approach effectively improves cache hit rates and significantly reduces latency in vehicular communication environments, yielding superior performance.

In this study, we propose a mobile edge computing (MEC)-enabled vehicular Metaverse system designed for augmented reality (AR) services, where vehicles on the road can access the Metaverse service through nearby Metaverse service providers (MSPs) equipped with MEC servers. In this system, vehicles are charged for their use of computational and communication resources. Due to varying positions and viewing angles, vehicles may have different content preferences. To minimize cost while ensuring optimal quality of experience (QoE), we formulate an optimization problem that adjusts content resolution and resource allocation to match individual vehicle needs. To address this optimization problem, we introduce a deep reinforcement learning (DRL) algorithm integrated with active inference theory to solve the decision-making problem with the performance of low latency and high efficiency. Simulation results demonstrate that our proposed scheme outperforms comparative algorithms in comprehensive performance, providing an effective solution for optimizing AR-enabled vehicular Metaverse systems.

With the rapid expansion of smart applications in 6G-enabled low-altitude economy networks, edge intelligence has emerged as an essential technology to reduce latency and energy consumption by dynamically deploying edge servers close to end users. Unmanned aerial vehicles (UAVs) can each carry a small-sized edge server to offer computation services to multiple end users. In this paper, we consider a non-orthogonal multiple access (NOMA)-based two-layer UAV swarm edge intelligence network, which includes task-generating UAVs (TUAVs) and an edge server UAV (SUAV). SUAV offers extra computation capacity to all TUAVs to process their computational tasks remotely. To minimise the long-term total service delay and the total energy consumption of all TUAVs based on the tasks’ transmission and processing, we propose a proximal policy optimisation (PPO)-based offloading decision and resource allocation joint optimisation algorithm by optimising offloading decisions, communication and computation resources allocation, and transmit power for all TUAVs. Simulations demonstrate that our approach outperforms benchmark solutions in terms of long-term total service delay and total energy consumption of all TUAVs in a UAV swarm system.

With the proliferation of vehicular user equipment (V-UE) in the Internet-of-Vehicles (IoV) systems, cloud computing alone cannot process all V-UE tasks, especially those latency-sensitive ones. Although static roadside fog nodes have been employed to offload computation from V-UEs, mobile fog nodes carried by vehicles that have the potential to further improve the performance of computation offloading for vehicular tasks have not been sufficiently studied for IoV systems. In this paper, we consider a mixed cloud/vehicular-fog computing (VFC) system that employs vehicle-carried fog nodes (V-FNs) in addition to cloud servers to offload tasks from V-UEs. To minimise the maximum service delay (which includes the transmission delay, queueing delay, and processing delay) among all V-UEs, we jointly optimise the offloading decisions of all V-UEs, the computation resource allocation at all V-FNs, the allocation of resource block (RB) and transmission power for all V-UEs while considering the mobility of V-UEs and V-FNs. The joint optimisation is solved by devising a fireworks algorithm-based offloading decision optimisation scheme, in conjunction with a bisection method-based V-FN computation resource allocation scheme and a clustering-based communication resource allocation scheme. Simulation results show that our proposed schemes outperform the benchmarks in terms of service the maximum delay among all V-UEs.

The orthogonal time frequency space (OTFS) modulation has garnered significant attention due to its potential to combat the frequency Doppler effect in high-mobility scenarios, especially in the low earth orbit (LEO) satellite communication. However, facilitating the OTFS in multiple-input and multipleoutput (MIMO) satellite communication requires accurate channel state information, which is a challenging task. To this end, we propose a novel deep-learning-based framework to enhance the channel estimation accuracy in a MIMO satellite communication network by exploiting the channel correlation of the OTFS-MIMO channels. Through extensive simulations, our proposed framework shows an impressive capability to predict MIMO-OTFS channels with remarkable accuracy enhancement. The effect of dataset and data distribution on the channel estimation accuracy and generalization are also revealed.

With the development of Fifth Generation (5G)/Sixth Generation (6G) -enabled Internet of Things (IoT) networks, different user equipment (UE) dynamically generates massive raw data and delay-sensitive computation tasks to be offloaded and processed at the mobile edge computing (MEC) nodes. In this paper, we propose a comprehensive digital twin-empowered UAV-assisted edge intelligent IoT framework, which enables UEs to offload their delay-sensitive tasks to a UAV-assisted MEC node. We aim to minimise the maximum total service delay including the transmission delay and the processing delay among all UEs. A deep deterministic policy gradient-based offloading and resource allocation optimisation algorithm, named (DDPG-ORAO), is proposed to optimise task offloading decisions among all UEs, which jointly optimising the communication and computation resources allocation among all UEs and all UAV-assisted MEC nodes. Simulation results show that our proposed optimisation algorithm outperforms the benchmarks in terms of the total service delay of all UEs.

Vehicular communications in future sixth-generation (6G) networks are expected to leverage integrated sensing and communications (ISACs) and mobile edge computing (MEC) techniques. However, the rapid proliferation of vehicle user equipment (V-UE) and the diversity of ISAC-aided and MEC-empowered vehicular communication and computation services demand a more intelligent and efficient resource allocation framework for the next-generation vehicular networks. To address this issue, we propose a comprehensive ISAC-aided vehicle-to-everything (V2X) MEC framework, where the V-UEs can offload their tasks to the edge server collocated at the roadside unit (RSU). We aim to minimize the long-term average total service delay of all the V-UEs by jointly optimizing the offloading decisions of all the V-UEs, the computation resource allocation at the ISAC-aided RSU, the transmission power, and the allocation of resource blocks for all the V-UEs, where the total service delay of a V-UE includes the task processing delay and the transmission delay if the V-UE offloads its task to the RSU. To solve the formulated mixed integer nonlinear programming problem, we design a multiagent deep deterministic policy gradient (MADDPG)-based offloading optimization and resource allocation algorithm (MADDPG-O2RA2). Simulation results demonstrate that our proposed algorithm outperforms the benchmarks in terms of convergence and the long-term average delay among all the V-UEs.

In order to support delay-sensitive applications of healthcare internet of things (HIoT) network systems, it is necessary to allow e-health wearable devices to process their computational applications at an artificial intelligence (AI) based mobile edge computing (MEC) server. Existing works mainly focused on optimising the offloading decisions of all users while minimising their corresponding energy consumption and/or transmission/processing delays. However, the huge data volumes offloading and/or privacy-preserving data processing have been largely ignored. In this paper, we consider an edge intelligence-based e-health wireless sensor network system, where the queues at the MEC server follow M/M/1 queueing model. To minimise energy consumption among all e-health wearable devices, we propose to jointly optimise the FL model task offloading decisions of all e-health wearable devices while guaranteeing delay constraints (which include the transmission delay, queueing delay and processing delay) and global/local FL training accuracy. This is achieved by devising an FL-based edge intelligence queueing offloading decision optimisation algorithm. Simulation results demonstrate that our proposed algorithm achieves a much lower maximum service delay than the benchmarks.

In order to support delay-sensitive applications of vehicle equipment (V-UE) in the Internet-of-Vehicles (IoV) systems, it is necessary to allow V-UEs to offload their computationally intensive applications to a cloud or fog computing server. Existing works mainly focused on minimising the transmission and processing delays while ignoring the mobility of V-UEs and/or the queueing delays at the cloud or fog servers. In this paper, we consider a vehicular network supported by a mixed fog and cloud computing system, where the queues at the fog node (FN) and the cloud centre are modelled following the M/M/1 and M/M/C queueing models, respectively. To minimise the maximum service delay (which includes the transmission delay, queueing delay and processing delay) among the V-UEs, we propose to jointly optimise the offloading decisions of all V-UEs and the computation resource allocation at the FN while considering the V-UEs’ mobility and queueing delays at the FN and cloud centre. This is achieved by devising a fireworks algorithm-based offloading decision optimisation algorithm in conjunction with a bisection method-based FN computation resource allocation scheme. Simulation results demonstrate that our proposed algorithm achieves a much lower maximum service delay than the benchmarks.

With the ever-increasing demand for wireless connections from vehicles, vehicle-to-everything (V2X) communication has become an emerging technology to enable vehicles to communicate with other vehicles, pedestrians, and communication infrastructures. In the meantime, most communication demands are initiated by human users, whose social-activities and attributes would influence their communication demands and requirements. However, the existing resource allocation schemes for V2X communications mainly focus on the physical domain, and have largely ignored the influence of the social domain. In this paper, we propose a social-aware clustering resource allocation (SACRA) algorithm to maximise the sum vehicle-to-infrastructure (V2I) capacity while guaranteeing the reliability of all vehicle-to-vehicle (V2V) links. In SACRA, firstly, we classify all vehicles into different social communities according to their social attributes. Secondly, in each social community, the V2I links are allocated with orthogonal spectrum resources, and the V2V links are divided into different clusters, each sharing the spectrum resource of a distinct V2I link, via a graph partitioning algorithm that minimises the intra-cluster interference. Compared with non-social-aware algorithms, simulation results demonstrate that the sum V2I capacity is improved by our proposed SACRA.